Radiomics analysis based on 18F-fluorodeoxyglucose positron emission tomography/computed tomography for differentiating the histological classification of peripheral neuroblastic tumours

Abstract

Aim

To create a radiomics nomogram based on ¹⁸F-fluorodeoxyglucose positron emission tomography/computed tomography and assess the efficacy of this nomogram in differentiating ganglioneuroblastoma from neuroblastoma in peripheral neuroblastic tumours (PNTs).

Materials and Methods

One hundred and ninety-nine patients with PNTs were retrospectively included, including 115 neuroblastoma patients and 84 ganglioneuroblastoma patients, who were randomly split into the training and test sets according to a ratio of 7:3. The 3D slicer was used to delineate the primary tumour, then radiomics features were extracted and selected, and a radiomics model was built using the optimal radiomics features. The clinical model was constructed from independent clinical risk factors. A radiomics nomogram was developed by multivariate logistic regression analysis incorporating independent clinical risk factors and radiomics features. Model performance was assessed using receiver operating characteristic curves and decision curve analysis.

Results

The radiomics model based on the selection of 14 radiomics features was developed. The clinical model was constructed by combining age at diagnosis and 1p aberrations. The radiomics nomogram demonstrated the optimal diagnostic value in distinguishing between ganglioneuroblastoma and neuroblastoma, with an area under the receiver operating characteristic curve of 0.857 and 0.795 in the training and test sets, respectively. The decision curve analysis and calibration curves also showed good performance for the nomogram.

Conclusions

The radiomics nomogram could improve diagnostic performance by differentiating the histological classification of PNTs.